The acute effects of morphine upon social behavior were investigated using the three-chamber test for sociability in mice, as previously reported (Piccin et al., 2022). During the habituation phase of the test, male C57BL/6J mice spent similar time in the regions of interests (ROIs, side half-chambers) of the apparatus (Fig. 1A-B and Table S2). Analysis of the sociability phase revealed a pretreatment X treatment X repeated measures interaction effect (Table S2). Unlike saline-treated mice, vehicle/morphine-treated mice spent similar time in the ROIs containing the unfamiliar conspecific or the object (P=0.823), indicating sociability deficits (Fig. 1C). In contrast, antalarmin/morphine-treated mice spent more time with the unfamiliar conspecific than with the object (P<0.005), indicating unaltered sociability (Fig. 1C). Accordingly, analysis of sociability ratio revealed a pretreatment effect (F_1,32=11.598, P<0.005), a treatment effect (F_1,32=4.713, P<0.05) and a pretreatment X treatment interaction effect (F_1,32=8.718, P<0.01). Vehicle/morphine-treated mice showed lower sociability ratio than vehicle/saline-treated mice (P<0.005, Fig. 1D). In contrast, antalarmin/morphine-treated mice did not differ from saline-treated mice (P=0.661) and showed higher sociability ratio than vehicle/morphine-treated mice (P<0.005, Fig. 1D). During the habituation phase, morphine-treated female C57BL/6J mice spent less time in the ROIs of the apparatus than saline-treated mice (P<0.0005), independently of vehicle or antalarmin pretreatment (Fig. 1E and Table S2). Moreover, analysis of the sociability phase revealed a treatment X repeated measures interaction effect but no pretreatment X treatment X repeated measures interaction effect (Table S2). Indeed, unlike saline-treated mice, morphine-treated mice spent similar time in the ROIs containing the unfamiliar conspecific or the object (P=0.259), independently of vehicle or antalarmin pretreatment (Fig. 1F). Accordingly, analysis of sociability ratio revealed no pretreatment effect (F_1,21=0.035, P=0.852), a treatment effect (F_1,21=8.698, P<0.01) but no pretreatment X treatment interaction effect (F_1,21=0.018, P=0.894). Morphine-treated mice showed lower sociability ratio than saline-treated mice (P<0.05), independently of vehicle or antalarmin pretreatment (Fig. 1G). During the three-chamber test, morphine-treated male, but not female, mice travelled more distance than saline-treated mice (P<0.05), independently of vehicle or antalarmin pretreatment (Table S3 and Fig. S1A-B). Also, overall mice travelled more distance during the habituation than during the sociability phase (P<0.05), indicating familiarization with the test apparatus (Table S3 and Fig. S1A-B). As detailed below in the Materials and methods section, we did not statistically compare the sexes and so are limited in making direct conclusions about sex differences. However, the present results indicate a sex-linked role for the CRF₁ receptor in social behavior deficits induced by morphine. Moreover, morphine effects upon sociability seemed unrelated to locomotor activity.

To investigate the neural substrates of CRF₁ receptor-mediated sociability deficits induced by morphine, electrophysiology studies examined firing frequency of PVN neurons (Fig. 2A). In male C57BL/6J mice, analysis of firing frequency of all of the recorded cells (n=110) revealed a pretreatment effect (F_1,106=7.894, P<0.01), a treatment effect (F_1,106=13.350, P<0.0005) and a pre-treatment X treatment interaction effect (F_1,106=4.208, P<0.05). Vehicle/morphine-treated mice showed higher firing frequency than vehicle/saline-treated mice (P<0.0005, Fig. 2B). In contrast, antalarmin/morphine-treated mice did not differ from saline-treated mice (P=0.552) and showed lower firing frequency than vehicle/morphine-treated mice (P<0.005, Fig. 2B). On the other hand, analysis of firing frequency of all of the recorded cells (n=93) in female C57BL/6J mice revealed no pretreatment effect (F_1,89=0.049, P=0.826), a treatment effect (F_1,89=20.476, P<0.0001) but no pre-treatment X treatment interaction effect (F_1,89=1.045, P=0.310). Morphine similarly increased firing frequency in vehicle- or antalarmin-pretreated mice, as compared to saline-treated mice (P<0.0005, Fig. 2D). These results indicate a critical role for the CRF₁ receptor in PVN neuronal activity induced by morphine in male, but not in female, mice. Notably, the sex-dependent effects of antalarmin upon neuronal firing closely mimicked the social behavior results, indicating a link between PVN activity and sociability deficits induced by morphine.

The specific role for the CRF₁ receptor in morphine-induced sociability deficits was further investigated using the genetic mouse model of CRF₁ receptor-deficiency. We first tested n=3 CRF₁+/+ and n=3 CRF₁+/− male mice using the same morphine dose (2.5 mg/kg) employed in the C57BL/6J mice. However, during the whole 10-min habituation phase of the three-chamber test, all of the six morphine-treated animals remained in the central chamber of the apparatus, suggesting that the morphine dose used was relatively high. Thus, we decided to use a substantially lower morphine dose, i.e., 0.625 mg/kg (Fig. 3A). During the habituation phase, morphine (0.625 mg/kg) reduced the time spent in both ROIs of the three-chamber apparatus in CRF₁+/− (P<0.05 vs. saline-treated CRF₁+/− mice), but not in CRF₁+/+ or CRF₁−/−, male mice (Fig. 3B and Table S4). Analysis of the sociability phase revealed a genotype X treatment X repeated measures interaction effect (Table S4). Unlike saline-treated mice, morphine-treated CRF₁+/+ and CRF₁+/− mice spent similar time in the ROIs containing the unfamiliar conspecific or the object (P=0.873), indicating sociability deficits (Fig. 3C). In contrast, morphine-treated CRF₁−/−mice spent more time with the conspecific than with the object (P<0.005), indicating unaltered sociability (Fig. 3C). Accordingly, analysis of sociability ratio revealed no genotype effect (F_2,58=2.641, P=0.080), a treatment effect (F_1,58=7.478, P<0.01) and a genotype X treatment interaction effect (F_2,58=4.994, P<0.01). Morphine-treated CRF₁+/+ mice showed lower sociability ratio than saline-treated CRF₁+/+ mice (P<0.05, Fig. 3D). In contrast, morphine-treated CRF₁−/−mice did not differ from saline-treated mice (P=0.819) and showed higher sociability ratio than morphine-treated CRF₁+/+ and CRF₁+/− mice (P<0.05, Fig. 3D). Interestingly, unlike CRF₁+/+ and CRF₁−/− mice, morphine-treated CRF₁+/− mice almost differed from saline-treated CRF₁+/− mice (P=0.065), suggesting a gene expression-dependent effect of CRF₁ receptor-deficiency (Fig. 3D). During the three-chamber test, overall morphine did not affect distance travelled (Table S4). Moreover, saline-treated mice and morphine-treated CRF₁+/+, but not CRF₁+/− or CRF₁−/−, mice travelled more distance during the habituation than during the sociability phase (P<0.05, Fig. S2 and Table S4), further indicating dissociation between the locomotor and the sociability effects of morphine. Thus, the similar results obtained with CRF₁−/− and antalarmin-treated C57BL/6J male mice strengthened the notion of a key role for the CRF₁ receptor in sociability deficits induced by morphine.

We then assessed the effect of morphine (0.625 mg/kg) in female CRF₁ receptor-deficient mice. However, as shown in Table S5, during the habituation phase of the three-chamber test only 2/8 CRF₁+/+ mice treated with saline and 2/8 CRF₁+/+ mice treated with morphine visited both side chambers of the apparatus. Also, despite all saline-treated CRF₁+/− mice (n=4) visited both side chambers of the apparatus, this occurred only in 3/8 CRF₁+/− mice treated with morphine. Thus, we could not obtain a reliable amount of data using a reasonable number of female mice, at least under our experimental conditions and with the 0.625 mg/kg morphine dose.

Analysis of firing frequency of PVN neurons in male CRF₁ receptor-deficient mice revealed a genotype effect (F_2,119=8.498, P<0.0005), a treatment effect (F_1,119=31.816, P<0.0001) and a genotype X treatment interaction effect (F_2,119=7.224, P<0.005). Morphine (0.625 mg/kg) increased firing frequency in CRF₁+/+ (P<0.0005) and in CRF₁+/− (P<0.005), but not in CRF₁−/− (P=0.987), mice, as compared to same-genotype saline-treated mice (Fig. 3E). Notably, morphine-treated CRF₁+/− mice showed lower or higher firing frequency than morphine-treated CRF₁+/+ (P<0.05) or CRF₁−/− (P<0.005) mice, respectively, indicating a CRF₁ gene expression-dependent effect (Fig. 3E). These results further support an essential role for the CRF₁ receptor in PVN neuronal firing induced by morphine. Moreover, the lack of morphine effects upon neuronal firing and sociability in CRF₁−/− mice indicates once more a link between PVN activity and social behavior.

In male C57BL/6J mice, 40 cells expressed both OXY and AVP, 49 cells expressed AVP but not OXY, 6 cells expressed OXY but not AVP and 15 cells expressed neither OXY nor AVP (Fig. 4B). Vehicle/morphine-treated mice showed higher firing frequency of OXY/AVP-expressing neurons than vehicle/saline-treated mice (P<0.0005; Fig. 4C and Table S6). In contrast, antalarmin/morphine-treated mice did not differ from saline-treated mice (P=0.782) and showed lower firing frequency than vehicle/morphine-treated mice (P<0.0005, Fig. 4C). On the other hand, morphine increased firing frequency of neurons expressing AVP, but not OXY, independently of vehicle or antalarmin pretreatment (P<0.05; Fig. 4D and Table S6). In female C57BL/6J mice, 31 cells expressed both OXY and AVP, 38 cells expressed OXY but not AVP, 7 cells expressed AVP but not OXY and 17 cells expressed neither OXY nor AVP (Fig. 4E). Morphine increased firing frequency of neurons co-expressing OXY and AVP, independently of vehicle or antalarmin pretreatment (P<0.005; Fig. 4F and Table S6). Similarly, morphine increased firing frequency of neurons expressing OXY, but not AVP, independently of vehicle or antalarmin pretreatment (P<0.005; Fig. 4G and Table S6). These results indicate a sex-specific role for the CRF₁ receptor in morphine-induced firing of PVN OXY-expressing neurons, suggesting that CRF modulates brain OXY responses to substances of abuse.

Male and female C57BL/6J mice were bred in-house and derived from mice originally purchased from Janvier Labs (Le Genest-Saint-Isle, France). Male and female CRF₁+/+, CRF₁+/− and CRF₁−/− mice previously generated on a mixed C57BL/6Jx129 background were bred in-house from mating CRF₁+/− mice and genotype identified by PCR analysis of tail DNA (Smith et al., 1998). The colony room (22±2°C, relative humidity: 50–60%) was maintained on a 12 h light/dark cycle (lights on at 08:00). Mice were housed in groups of 2-4 in transparent polycarbonate cages (29.5 x 11.5 x 13 cm, L x W x H) containing bedding and a cotton nestlet (SAFE, Augy, France) and were 12-28 weeks old at testing. Standard laboratory food (3.3 kcal/g; SAFE, Augy, France) and water were available ad libitum. All studies were conducted in accordance with the European Communities Council Directive 2010/63/EU, were approved by the local Animal Care and Use Committee and complied with the ARRIVE Guidelines (Kilkenny et al., 2010).

The three-chamber task allowed the study of the preference for an unfamiliar same-sex conspecific vs. an object and was carried out as previously reported (Piccin et al., 2022). The three-chamber apparatus was a rectangular box (60 x 40 x 20 cm, L x W x H) divided in three equal chambers and made of dark grey polypropylene. Dividing transparent Plexiglas walls had small squared doors (8 x 8 cm) that could be manually opened and closed. The central chamber was empty and each side chamber contained a round wire cage (12 cm diameter, 14 cm high, with bars spaced 1 cm apart) placed in one half-portion of the chamber. The three-chamber test was conducted during the light phase of the 12 h light/dark cycle and light intensity in the apparatus was ∼10 lux. The subject mice were handled 1 min/day during the three days preceding the three-chamber experiment. On the fourth day, C57BL/6J mice were treated with either vehicle or antalarmin (20 mg/kg) and returned to their home-cage. One hour later, they were treated with either saline or morphine (2.5 mg/kg) and immediately tested in the three-chamber task (Fig. 1A). CRF₁+/+, CRF₁+/− and CRF₁−/− mice were just treated with either saline or morphine (0.625 mg/kg) and immediately tested in the three-chamber task (Fig. 3A). The three-chamber test consisted of three phases: pre-habituation, habituation and sociability. During the pre-habituation phase, the subject mouse was confined to the central chamber for 5 min. Then, the doors were opened and the mouse could freely explore the three chambers and the empty wire cages for 10 min (habituation phase). During the subsequent 10 min, the subject mouse could freely explore the entire apparatus with one wire cage containing an unfamiliar same-sex mouse and the other an object, i.e., a plastic bottle cap (sociability phase). The unfamiliar mice were C57BL/6J mice sex- and age-matched with the subject mice. During the three days preceding testing, the unfamiliar mice were handled 1 min/day and habituated to the wire cages for 10 min/day, with the wire cage habituation taking place in the three-chamber apparatus on the second and third day. The position (left or right side chamber) of the unfamiliar mouse was counterbalanced within each experimental group. Between each tested mouse, the apparatus was cleaned with water and the wire cages with 70% ethanol and then water. Videos were acquired and analyzed with a home-made tracking system. In particular, time (s) spent by the tested mouse in the regions of interest (ROIs, side half-chambers) containing the wire cages was taken as a measure of sociability (Fig. 1A). Indeed, our prior studies reliably demonstrated that the latter measure positively correlated with the number of nose-to-nose contacts with the unfamiliar mouse (Piccin and Contarino, 2020b). Moreover, ratio of time spent in the ROI containing the unfamiliar mouse positively correlated with the ratio of time spent with the nose in the wire cage containing the unfamiliar mouse (Piccin and Contarino, 2020b). Moreover, to control for locomotor activity, distance (m) travelled throughout the whole apparatus during the habituation and the sociability phases of the test was examined. Sociability ratio was calculated as percentage of time spent in the ROI containing the unfamiliar mouse over the total time spent in both ROIs containing the wire cages.

C57BL/6J mice were injected with either vehicle or antalarmin (20 mg/kg) and, one hour later, with either saline or morphine (2.5 mg/kg). CRF₁ receptor-deficient mice were just injected with either saline or morphine (0.625 mg/kg). Ten minutes after saline or morphine administration, mice were anesthetized by intraperitoneal (i.p.) injection of ketamine (100 mg/kg) / xylazine (10 mg/kg) until reflexes to tail- or toe-pinching were lost. Before brain removal, animals were intracardially perfused with an ice-cold bubbled (95% O₂ / 5% CO₂) sucrose-based saline solution containing (in mM): NaH₂PO₄ 1.25, KCl 2.5, CaCl₂ 0.5, MgSO₄ 10, D-glucose 10, NaHCPO₃ 26. Brain tissue was rapidly removed and 300 μm coronal slices containing the PVN were cut using a vibroslicer (Leica VT100S, Leica Biosystems, Germany). Slices were then allowed to recover for at least 1 h at 30°C in a holding chamber filled with oxygenated (95 % O₂ / 5 % CO₂) artificial cerebrospinal fluid (aCSF) composed of (in mM): NaCl 126, KCl 2.5, CaCl₂ 2, MgSO₄ 2, NaH₂PO₄ 1.25, NaHCO₃ 26, glucose 10 (pH 7.3, 290 mOsm).

Cell-attached patch-clamp recordings from PVN neurons were made at room temperature in current clamp conditions under continuous perfusion of oxygenated aCSF composed of (in mM): NaCl 126, KCl 3, CaCl₂ 1.6, MgSO₄ 1.5, NaH₂PO₄ 1.25, NaHCO₃ 26, glucose 10. Throughout recordings, GABAergic and glutamatergic inputs were blocked with gabazine (1 _μM) and 10μM of the NMDA and non-NMDA receptor antagonists 6,7-dinitroquinoxaline-2,3(1H,4H) dione (DNQX) and D(-)-2-amino5-phosphonopentanoic acid (AP5), respectively. Neurons were visualized with an upright Nikon Eclipse FN1 microscope (Nikon, Japan) with infrared illumination. Recording borosilicate electrodes were filled with an internal solution containing K-Gluconate 120 mM, KCl 20 mM, MgCl₂ 1.3 mM, EGTA 1 mM, HEPES 10 mM, CaCl₂ 0.1 mM, GTP 0.03 mM, cAMP 0.1 mM, leupeptine 0.01 mM, D-Mannitol 77 mM and Na 2 ATP 3 mM (pH 7.3). Moreover, biocytin 0.1% was added to the internal solution in order to post-visualize recorded neurons. Data were collected online with a Multiclamp 700B amplifier (Molecular Devices, USA) and acquired with Axograph X software (Axograph, Australia). Electrophysiology recordings were analyzed offline using the Axograph X software.

The phenotype of the patched and recorded cells was assessed by immunohistochemistry. After electrophysiological recording, slices were fixed with 4% paraformaldehyde overnight at 4°C. Biocytin was then revealed with FITC-Streptavidin (1/300, Vector Laboratories). OXY and AVP immunohistochemical labeling was performed at the same time using as first antibodies mouse anti-OXY monoclonal antibody (1/1000, Millipore MAB5296) and T-5048 Guinea pig anti (Arg⁸)-vasopressin antibody (1/1000, BMA biomedicals). Alexa fluor 488 goat anti-mouse IgG (1/500, Life technology) and Alexa fluor 647 donkey anti-guinea pig IgG (1/500, Life technology) were used as secondary antibodies. Immunostainings were acquired using a confocal Zeiss LSM900 microscope. Serial optical sections were obtained at a Z-step of 1.2 μm and imaged using an objective 10X or 20X /1.00 numerical aperture.

Antalarmin hydrochloride (20 mg/kg; TOCRIS, Lille, France) was dissolved in acidified saline (pH ∼2.5) and injected per os (p.o.) by gavage. Morphine hydrochloride (0.625 or 2.5 mg/kg; Francopia, Gentilly, France) was dissolved in physiological saline and injected i.p. Control mice were injected p.o. or i.p. with the appropriate vehicle (acidified or physiological saline) and volume of administration was always 10 ml/kg.

Each mouse was assigned a unique identification number that was used to conduct blind testing and data analysis. To prevent strong initial preferences from biasing the three-chamber sociability results, animals exploring each ROI containing the wire cage for more than 80% (or less than 20%) of the total time spent in both ROIs during the habituation phase (10 min) were excluded from data analysis. The number of animals excluded within each experimental group is reported in Table S1A-B. For simplification and illustration purposes, we did not statistically compare the sexes in the behavior and electrophysiology studies. Thus, within each sex, the three-way repeated measures analysis of variance (ANOVA) with pre-treatment (vehicle vs. antalarmin) and treatment (saline vs. morphine) as between-subjects factors and side (mouse vs. object) or test phase (habituation vs. sociability) as a within-subject factor was used to analyze time spent in the ROIs or distance travelled during the three-chamber test by C57BL/6J mice. A three-way repeated measures ANOVA with genotype (CRF₁+/+ vs. CRF₁+/− vs. CRF₁−/−) and treatment (saline vs. morphine) as between-subjects factors and side (mouse vs. object) or test phase (habituation vs. sociability) as a within-subject factor was used to analyze time spent in the ROIs or distance travelled during the three-chamber test by CRF₁ receptor-deficient mice. The two-way ANOVA with pre-treatment (vehicle vs. antalarmin) or genotype (CRF₁+/+ vs. CRF₁+/− vs. CRF₁−/−) and treatment (saline vs. morphine) as between-subjects factors was used to analyze sociability ratio and the firing frequency (Hz) results of the electrophysiology studies. The accepted value for significance was P<0.05. Following significant interaction effects, the Newman-Keuls post-hoc test was used for individual group comparisons. Statistical analyses were performed using the Statistica software (Version 10). Data graphs were created using GraphPad Prism and Adobe Illustrator.